Mac preview
NativeRig for macOS
Apple silicon · macOS 14+
A complete three-platform preview release has not been published yet.
Tested end to end on Apple silicon. This build is unsigned, and broader Mac testing is still in progress.
Technical preview
Max out your AI workstation.
NativeRig easily and securely chooses the latest and best AI model for your AI workstation’s specs.
Use what you paid for. Runtime settings are derived from your accelerator, memory, storage, and local measurements.
Skip the model lottery. A recommendation must still pass speed, memory, response, and coding-tool checks on your machine.
Get a working result. NativeRig owns setup, validation, start, stop, upgrade, and rollback—not just the model download.
The value
Turn workstation specifications into a proven coding setup.
A model that fits in memory can still be slow, fail coding tools, or leave expensive hardware underused. NativeRig combines a current recommendation with measurements on your machine, then activates only a setup that passes locally.
- 1 · Match
Start from your hardware, not model hype.
NativeRig inspects the operating system, accelerator, memory, and free storage, then shows an installable coding candidate and its download size when available. Validate one by default, or compare up to three.
- 2 · Tune
Convert available capacity into useful performance.
NativeRig derives context size, GPU offload, threads, batch size, and memory reserve, then measures response speed, memory use, and coding-tool reliability. Failed candidates never replace the working setup.
- 3 · Work
Finish with a working agent, not a running server.
NativeRig requires a real model response and coding-tool check before success. The current preview then opens your project in managed OpenCode while leaving existing global and project settings unchanged.
Download
Download a tester build for your computer.
These are pre-release builds, not generally supported installers. Choose your operating system to see the exact hardware evidence and remaining qualification work.
Initial setup needs internet access and may download tens of gigabytes. NativeRig asks first and shows the model, source, and size when available. After setup, the coding profile runs offline by default.
For current testers
Finished a NativeRig test run?
The feedback form is on a separate page. Bring the CSV NativeRig saved, then report whether installation and first launch worked.